linux-xfs.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Darrick J. Wong" <djwong@kernel.org>
To: djwong@kernel.org, guaneryu@gmail.com
Cc: Matthew Wilcox <willy@infradead.org>,
	linux-xfs@vger.kernel.org, fstests@vger.kernel.org, guan@eryu.me
Subject: [PATCH 3/4] xfs/420: fix occasional test failures due to pagecache readahead
Date: Tue, 15 Mar 2022 20:30:20 -0700	[thread overview]
Message-ID: <164740142033.3371628.11850774504699213977.stgit@magnolia> (raw)
In-Reply-To: <164740140348.3371628.12967562090320741592.stgit@magnolia>

From: Darrick J. Wong <djwong@kernel.org>

Every now and then, this test fails with this golden output:

--- xfs/420.out
+++ xfs/420.out.bad
@@ -29,7 +29,7 @@
 Whence Result
 DATA   0
 HOLE   131072
-DATA   196608
+DATA   192512
 HOLE   262144
 Compare files
 c2803804acc9936eef8aab42c119bfac  SCRATCH_MNT/test-420/file1

Curiously, the file checksums always match, and it's not *forbidden* for
the page cache to have a page backing an unwritten extent that hasn't
been written.

The condition that this test cares about is that block 3 (192k-256k) are
reported by SEEK_DATA as data even if the data fork has a hole and the
COW fork has an unwritten extent.  Matthew Wilcox thinks this is a side
effect of readahead.

To fix this occasional false failure, call SEEK_DATA and SEEK_HOLE only
on the offsets that we care about.

Suggested-by: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 tests/xfs/420 |   33 +++++++++++++++++++++------------
 1 file changed, 21 insertions(+), 12 deletions(-)


diff --git a/tests/xfs/420 b/tests/xfs/420
index 12b17588..d38772c9 100755
--- a/tests/xfs/420
+++ b/tests/xfs/420
@@ -50,6 +50,24 @@ _scratch_mount >> $seqres.full 2>&1
 testdir=$SCRATCH_MNT/test-$seq
 mkdir $testdir
 
+# pagecache readahead can sometimes cause extra pages to be inserted into the
+# file mapping where we have an unwritten extent in the COW fork.  Call lseek
+# on each $blksz offset that interests us (as opposed to the whole file) so
+# that these extra pages are not disclosed.
+#
+# The important thing we're testing is that SEEK_DATA reports block 3 as data
+# when the COW fork has an unwritten mapping and the data fork has a hole.
+exercise_lseek() {
+	echo "Seek holes and data in file1"
+	$XFS_IO_PROG -c "seek -d 0" $testdir/file1
+	$XFS_IO_PROG -c "seek -h $((2 * blksz))" $testdir/file1 | sed -e '/Whence/d'
+	echo "Seek holes and data in file2"
+	$XFS_IO_PROG -c "seek -d 0" $testdir/file2
+	$XFS_IO_PROG -c "seek -h $((2 * blksz))" $testdir/file2 | sed -e '/Whence/d'
+	$XFS_IO_PROG -c "seek -d $((3 * blksz))" $testdir/file2 | sed -e '/Whence/d'
+	$XFS_IO_PROG -c "seek -h $((4 * blksz))" $testdir/file2 | sed -e '/Whence/d'
+}
+
 blksz=65536
 nr=8
 filesize=$((blksz * nr))
@@ -83,10 +101,7 @@ $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file1 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file2 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file3 >> $seqres.full 2>&1
 
-echo "Seek holes and data in file1"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file1
-echo "Seek holes and data in file2"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file2
+exercise_lseek
 
 echo "Compare files"
 md5sum $testdir/file1 | _filter_scratch
@@ -102,10 +117,7 @@ $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file1 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file2 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file3 >> $seqres.full 2>&1
 
-echo "Seek holes and data in file1"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file1
-echo "Seek holes and data in file2"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file2
+exercise_lseek
 
 echo "Compare files"
 md5sum $testdir/file1 | _filter_scratch
@@ -121,10 +133,7 @@ $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file1 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file2 >> $seqres.full 2>&1
 $XFS_IO_PROG -c "bmap -ev" -c "bmap -cv" $testdir/file3 >> $seqres.full 2>&1
 
-echo "Seek holes and data in file1"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file1
-echo "Seek holes and data in file2"
-$XFS_IO_PROG -c "seek -a -r 0" $testdir/file2
+exercise_lseek
 
 echo "Compare files"
 md5sum $testdir/file1 | _filter_scratch


  parent reply	other threads:[~2022-03-16  3:30 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-16  3:30 [PATCHSET 0/4] fstests: random fixes Darrick J. Wong
2022-03-16  3:30 ` [PATCH 1/4] generic/459: ensure that the lvm devices have been created Darrick J. Wong
2022-03-22  4:50   ` Zorro Lang
2022-03-16  3:30 ` [PATCH 2/4] common/xfs: fix broken code in _check_xfs_filesystem Darrick J. Wong
2022-03-22  4:57   ` Zorro Lang
2022-03-16  3:30 ` Darrick J. Wong [this message]
2022-03-22  5:13   ` [PATCH 3/4] xfs/420: fix occasional test failures due to pagecache readahead Zorro Lang
2022-03-22 16:10     ` Darrick J. Wong
2022-03-16  3:30 ` [PATCH 4/4] generic/673: fix golden output to reflect vfs setgid behavior Darrick J. Wong
2022-03-22  5:24   ` Zorro Lang
2022-03-16 22:12 ` [PATCH 5/4] xfs/076: only create files on the data device Darrick J. Wong
2022-03-22  8:31   ` Zorro Lang
2022-03-16 22:13 ` [PATCH 6/4] xfs/17[035]: fix intermittent failures when filesystem metadata gets large Darrick J. Wong
2022-03-22  7:15   ` Zorro Lang
2022-03-22 16:11     ` Darrick J. Wong

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=164740142033.3371628.11850774504699213977.stgit@magnolia \
    --to=djwong@kernel.org \
    --cc=fstests@vger.kernel.org \
    --cc=guan@eryu.me \
    --cc=guaneryu@gmail.com \
    --cc=linux-xfs@vger.kernel.org \
    --cc=willy@infradead.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).